Dr. Nasser Razack is currently engaged in an Interventional
Neuroradiology fellowship at The University of Virginia,
Charlottesville, VA. Dr. Jonas Goldstein is the director of
Interventional Neuro-radiology at Ashville Radiology
Associates, Ashville, VA. Dr. Mary E. Jensen is an Associate
Professor of Radiology and Neurosurgery and Director of
Interventional Neuroradiology at The University of Virginia,
Charlottesville, VA.
Epistaxis is a common event encountered by 60% of the
population.
1
Fortunately, only 6% of this group requires medical attention.
There are multiple etiologies of nasal bleeding, including
hypertension, facial trauma, iatrogenic complications, hereditary
hemorrhagic telangiectasia (HHT), benign or malignant tumors,
vascular abnormalities of the internal or external carotid
arteries, and coagulopathies. Despite the wide spectrum of
etiology, most cases of epistaxis are idiopathic.
Sokoloff et al
2
introduced embolization as a method for the treatment of epistaxis
in 1974. Despite the ease and effectiveness of his treatment
regimen and the fact that endovascular treatment has been shown
consistently to have equal or higher success rates with lower
morbidity and mortality than surgical ligation, endovascular
therapy has failed to be accepted as the primary treatment modality
for epistaxis unresponsive to nonsurgical therapies. In addition,
proximal surgical ligation of regional blood supply blocks access
and stimulates the recruitment of endovascularly inaccessible
and/or dangerous anastomoses, making future embolization risky or
impossible.
Patients with epistaxis that is unresponsive to medical,
topical, or direct-pressure therapy or patients with emergent,
life-threatening hemorrhage are considered candidates for
endovascular therapy. Important clinical information obtained prior
to the examination includes: pertinent past history for trauma,
neoplasm, or surgery; previous episodes of epistaxis and therapy
received; and the presumed site and cause of the hemorrhage.
Hematologic and coagulation studies should be recent, and
hemoglobin and/or platelet deficiencies should be corrected.
Diagnostic angiography and embolization are performed in the
same session. High-resolution digital angiography is used in all
cases because of its variable frame rate filming, road-mapping, and
live subtraction capabilities, and near instant views of
angiographic runs. Angiographic analysis of the "vascular map" of
the nasal cavity evaluates the site of hemorrhage and the adjacent
normal tissue, determines the most appropriate vessels for
embolization, and searches for dangerous anastomoses.
The arterial anatomy to the nasal fossa consists of a dual blood
supply from branches of both the external (ECA) and internal (ICA)
carotid arteries (figure 1). The majority of the ECA supply is via
the internal maxillary (sphenopalatine and greater palatine
branches) and facial arteries. Anterior and posterior ethmoidal
arteries arise off the ophthalmic artery to supply the nasal fossa.
The major supply to the nasal fossa is via the lateral and medial
branches of the sphenopalatine artery.
Angiography of the external carotid artery may show a variety of
findings depending on the underlying cause of the bleeding. In
hereditary hemorrhagic telangiectasia, angioectasia lends a typical
"corkscrew" appearance to the terminal branches of the internal
maxillary and/or facial arteries. Traumatic or iatrogenic injury
may show vascular irregularity, arterial transection,
pseudoaneurysm formation, or extravasation of contrast material.
Tumors such as juvenile nasoangiofibromas show dense staining of
the lesion with a persistent tumor blush. In patients with
hypertension or coagulopathies, as well as in idiopathic cases, the
vascular supply to the nasal mucosa is frequently normal in
appearance, and angiography is performed to rule out the presence
of other causes of epistaxis.
The order in which the vessels are studied depends on the
situation and the angiographer's preference. In patients who are
actively bleeding, the "most likely" vessel is evaluated first and
is treated. Then, adjacent vascular territories can be studied to
rule out other bleeding sites or contributing vessels. When the
source of hemorrhage is unknown and it is anticipated that
bilateral carotid branches will be embolized, the internal carotid
artery may be studied first to exclude an unexpected carotid siphon
or ophthalmic artery source. A global ECA study is then performed,
followed by superselective catheterization and embolization of the
targeted ECA branches. A postembolization common carotid injection
will ensure that internal carotid artery or ophthalmic feeders have
not recanalized the embolized territory.
Embolization is performed using a coaxial system consisting of a
larger guiding catheter, and a variable-stiffness microcatheter for
superselective cannulation. Many catheters, guidewires, and embolic
agents are currently available; this discussion is limited to the
more common materials used in the treatment of epistaxis.
Guiding catheters are required as part of the coaxial system
used in the catheterization of small vessels. Although these
catheters range in size from 4F to 10 F, most guides for epistaxis
embolization are 4F to 6F. The use of a smaller guiding catheter
decreases femoral artery compression time and complications but
requires road-mapping to be performed prior to the placement of the
microcatheter or through the microcatheter. Some of the newer 5F
and 6F guiding catheters have large inner lumens (0.054 to 0.066
inch) that allow road-mapping around the microcatheter and soft
tips that reduce the risk of vasospasm. Use of a femoral sheath
provides patient comfort and improves catheter movement and
response.
The tip of the guiding catheter is positioned in the parent
artery, where it acts as an introducer for the smaller
microcatheter, which passes through a rotating hemostatic valve
connected to the guiding catheter hub. The dead space is flushed
continuously with heparinized saline to prevent clot formation. A
three-way stopcock placed on the rotating valve sidearm allows the
injection of contrast material through the guiding catheter for
serial angiograms or road-mapping while the microcatheter is in
place.
In almost all cases, a steerable microcatheter is used in
combination with a microguidewire to access the targeted vessels.
The lumen of a steerable microcatheter is sufficient in size to
accept a variety of embolic agents, including particles, gelatin
sponge pledgets, platinum microcoils, and liquid agents. Steerable
microcatheters vary in size and stiffness and may have metallic
braiding in the wall, hydrophilic coating, and/or Teflon lining the
lumen. The choice of microcatheter is left to the operator's
preference, but braided, hydrophilic microcatheters often
demonstrate better "pushability" and tracking. Frequently, a small
45° angle is steamed into the end of the microcatheter to assist
advancement around curves or to engage vascular orifices. Steaming
should be done with a shaping mandril in place to prevent shrinkage
of the catheter tip. Flow-directed catheters are usually reserved
for treatment of high-flow lesions, such as arterio-venous
malformations or fistulae.
Like microcatheters, a variety of microguidewires exist and vary
in diameter size, stiffness, and materials used in construction.
Newer guidewires also have hydrophilic coatings to decrease
friction between it and the microcatheter. Larger diameter
microguidewires (0.014 to 0.016 inch) usually have better torque
control and provide a stiffer support for microcatheter advancement
than do smaller (0.010 inch) wires. Although these systems appear
to be relatively atraumatic, excessive guidewire or catheter
manipulation or wedging of the catheter tip in the artery can
induce vasospasm. Alteration of blood flow to the lesion may
prevent safe and effective embolization by increasing the risk of
reflux or opening collateral channels. If vasospasm occurs, it can
be relieved by the application of topical nitroglycerine
(Nitropaste, Fougera and Co., Melville, NY) or the administration
of intra-arterial vasodilators (i.e., papaverine). Removal of the
catheter and a "tincture of time" will also often alleviate the
vasospasm.
The embolic agents are chosen based upon the lesion being
treated, the type of occlusion desired (temporary versus
permanent), and the preference of the interventionalist.
Particulate agents are used in most cases of epistaxis, and create
mechanical blockage of the targeted vessels with individual
particles of uniform size and shape. Polyvinyl alcohol (PVA) is the
most common particulate agent and is supplied in a variety of size
ranges, from 45 to 1000 µm. This material adheres to vascular
endothelium, inducing endothelial proliferation and fibrosis. The
particles also incite thrombosis within the vascular bed, and
recanalization around the particles from collateral flow may occur.
Particles smaller than 150 to 200 µm are avoided because these are
more likely to pass through dangerous anastomoses, which may
occlude the vasa nervosum, resulting in cranial nerve palsies, or
may cause skin or muscular necrosis. Other particulate agents
include microfibrillar collagen and gelatin powder and/or shredded
sponge. Particulate size in gelatin powder ranges from 40 to 60 µm,
and, if chosen, must be used with extreme caution for the
above-cited reasons. Gelatin sponge can be shredded or rolled into
pledgets of 2 to 3 mm in length, for embolization.
Larger, more discrete agents include latex and silicone
balloons, stainless steel coils, and platinum microcoils. Balloon
use is most frequently seen in the treatment of fistulae or for
permanent parent artery occlusion. Microcoils may be used to
protect a normal vessel or to prevent accidental embolization
through a dangerous anastomosis. Use of microcoils as a proximal
occlusive agent in feeding vessels should be avoided in case
embolization is necessary. Liquid agents such as tissue adhesives
and absolute alcohol are reserved for specific situations and carry
significant risks. These materials should be used only by
experienced operators and very rarely have a place in the treatment
of epistaxis.
Treatment strategies vary depending on the etiology of
epistaxis. In patients with idiopathic epistaxis, angiographic
visualization of the hemorrhage site is rare. The approximate site
of the presumed bleeding should be determined by the clinician, and
the appropriate vascular territory is targeted for treatment. The
selected vessels are embolized as distally as possible, with
special attention to the presence of dangerous anastomoses between
branches of the internal and external carotid arteries. Patients
presenting with epistaxis from nasopharyngeal tumors are treated
with particles in a manner similar to idiopathic epistaxis.
The protocol followed at our institution involves particulate
embolization of the bilateral internal maxillary arteries, and one
or both facial arteries (figure 2A and B). The pterygopalatine
portion of the internal maxillary artery is catheterized
selectively to avoid embolization of the muscles of mastication,
which may result in trismus. In the facial artery, catheterization
is distal to the labial branches to prevent lip paresthesia. If the
facial artery cannot be negotiated, embolization is done proximal
to the labial artery with larger particles or gelatin sponge
pledgets. The blood supply to the nasal cavity is rich in
collateral circulation, and the risk of mucosal necrosis is low as
long as an appropriately sized embolic agent is selected. Small to
moderately sized (250 to 500 µm) PVA particles are commonly used
for most cases of epistaxis. To allow the operator to monitor
vessel runoff, stagnation of blood flow, and reflux of the embolic
agent, particles should be suspended and delivered in contrast
material. Embolization is complete when the smaller branches are no
longer visualized and the larger arteries take on a "pruned-tree"
appearance. A gelatin sponge pledget may be placed in the main
trunk following particulate embolization to promote hemostasis and
thrombosis (figure 2C). To assess results and the need for further
embolization, the nasal packing may be removed while the patient
remains on the table. We prefer to leave the packing in place
overnight, with subsequent removal by the clinician under
controlled conditions.
The internal carotid artery should be evaluated for the presence
of contributory feeders or other vascular abnormalities, such as
aneurysms, and, in cases of trauma, for evidence of vascular injury
(figure 3). Traumatic epistaxis may be associated with penetrating
injuries of the carotid artery or facial fractures. Small branch
dissections or pseudoaneurysms may be treated by permanent
occlusion of the feeding vessel using microcoils placed just
proximal to the site of vascular injury. Traumatic injury to the
internal carotid artery is considered an emergency and will most
likely result in occlusion of the parent artery.
The complication rate of transarterial particulate embolization
in experienced hands is <1%, although facial nerve paresis and
cerebral infarction have been reported. Complications are usually
caused by: selection of inappropriate embolic material; reflux of
embolic material caused by vascular spasm, insufficient selective
catheterization or overzealous injection; or failure to recognize
potentially dangerous anastomoses. Other complications include skin
necrosis when too forceful an injection of fine particles is
employed, and trismus with embolization of the muscles of
mastication. Some patients may complain of headache, sinus
tenderness, or low-grade fever. These postprocedural symptoms are
self-limiting and are treated easily with analgesics.
AR